Artificial intelligence is approximating human reasoning more and more closely all the time. Wide-scale adoption by business may be approaching, with important implications for how people live and work. AI is paving the way for new business models and raising questions about how people and machines can best work together. Now, thanks in part to cheaper and faster computing power, intelligent machines help doctors comb through troves of medical images to identify diseases early, allow manufacturers to predict when their machines will break (and fix them before that happens), and provide the "brains" behind increasingly autonomous vehicles. It's also playing a central role in the consumer market, powering the latest virtual assistant, for example, or the engine that matches Airbnb guests with the housing they want.

A swarm intelligence similar to the one that predicted Oscar winners and Kentucky Derby finishers has come to nearly unanimous conclusions about the presidential potential of Hilary Clinton and Donald Trump. From social issues to trustworthiness and ethics, the swarm spoke loud and clear, expressing practically the same sentiment for both candidates -- extreme pessimism. The swarm consisted of 85 Democratic, Republican, or independent American voters who were asked to answer identical questions on Clinton and Trump through the swarm intelligence platform UNU. The speed at which they came to a conclusion helps calculate the percentage of "brainpower" for a particular swarm. Anywhere between 70 and 85 people participated in each round.

Reuters reported Tuesday that Germany's Transport Minister Alexander Dobrindt plans legislation requiring self-driving cars be outfitted with data-recording systems in an effort to bring more accountability and to give automakers and engineers a chance to learn from the systems' mistakes. The proposal would enable autonomous car users to divert some of their attention from traffic while the systems are on as long as they are seated at the wheel but require that they maintain black boxes that track when the programs request drivers take over.

Classification is an important statistical learning tool. In real application, besides high prediction accuracy, it is often desirable to estimate class conditional probabilities for new observations. For traditional problems where the number of observations is large, there exist many well developed approaches. Recently, high dimensional low sample size problems are becoming increasingly popular. Margin-based classifiers, such as logistic regression, are well established methods in the literature. On the other hand, in terms of probability estimation, it is known that for binary classifiers, the commonly used methods tend to under-estimate the norm of the classification function. This can lead to biased probability estimation. Remedy approaches have been proposed in the literature. However, for the simultaneous multicategory classification framework, much less work has been done. We fill the gap in this paper. In particular, we give theoretical insights on why heavy regularization terms are often needed in high dimensional applications, and how this can lead to bias in probability estimation. To overcome this difficulty, we propose a new refit strategy for multicategory angle-based classifiers. Our new method only adds a small computation cost to the problem, and is able to attain prediction accuracy that is as good as the regular margin-based classifiers. On the other hand, the improvement of probability estimation can be very significant. Numerical results suggest that the new refit approach is highly competitive.

We consider the problem of predicting to which classes an image belongs, where the number of classes is large (many thousands or tens of thousands) and where each image typically belongs to multiple classes that should all be properly identified: multi-label, massively multi-class classification. In such classification problems, the best practice until now (for instance in use at Google, Inc.) has been to use a deep convolutional neural network such as the ones described in [19] or [18], culminating in a logistic regression layer with a sigmoid cross-entropy loss, with target labels encoded as high-dimensional sparse binary vectors. The use of logistic regression implies an important yet oft overlooked assumption made about the label space: the classes are considered to be statistically independent, each class being treated as an independent dimension in the label space. This is generally not the case in practice: mirroring statistical dependencies found in the real world, label spaces often have a well-defined internal structure, with some labels being more likely to cooccur than other labels. For instance, "sky" and "beach" are frequently cooccurring labels, while "crane" and "manta ray" are rarely cooccurring. The sigmoid cross-entropy loss with sparse binary targets does not allow to leverage such observations about the structure of the label space. 1 There is therefore an opportunity to exploit the internal structure of the label space for gains in training speed, precision, and recall. One simple way to achieve this is to project the labels onto a lower-dimensional manifold -an embedding space-where a distance function between embedded labels would capture useful statistical dependencies. An appropriate loss function may then allow a parametric model trained via stochastic gradient descent to benefit from the structure of the manifold during training and inference.

The following interview is one of many included in the report. As part of our ongoing series of interviews surveying the frontiers of machine intelligence, I recently interviewed Gurjeet Singh. Singh is CEO and co-founder of Ayasdi, a company that leverages machine intelligence software to automate and accelerate discovery of data insights. Author of numerous patents and publications in top mathematics and computer science journals, Singh has developed key mathematical and machine learning algorithms for topological data analysis. David Beyer: Let's get started by talking about your background and how you got to where you are today.

Bots are currently all the rage, but they're not a new concept. Bots are only as useful as the services they are integrated with, and their purpose is essentially automation -- that is, creating and executing actions based upon a set of criteria. In order to know where bots are going, though, we need to understand where they've been. While no one knows exactly when bots started, they're widely thought to have gotten off the ground with ELIZA. The bot was built by Joseph Weizenbaum, an MIT professor, in 1964.

A company that sells software that analyzes the human voice and touts the virtues of empathy, rapport and emotional intelligence is joining forces with West Point United States Military Academy in an effort to help cadets become better negotiators. Cogito Corp. is a Boston-based company that makes software that can analyze a person's voice in real-time. That information, the company says, can help customer service representatives show more empathy; the result is phone conversations that are more efficient and personalized, according to Cogito. Col. James Ness of West Point said that this kind of tech will help their students become better negotiators, a key skill for people in the military. "Cogito's behavioral analytics technology will systematically analyze communication patterns within negotiating sessions and provide insight into the cadet's psychological state," Ness, who directs the engineering psychology program at West Point, said in a statement.

Numerous companies in the healthcare space are experimenting with artificial intelligence, but what does the future hold in this sphere? You'll probably know IBM's supercomputer, Watson, from its 2011 appearance on Jeopardy. Up against two of the US quiz show's longest-running and highest-earning contestants, Watson clinched a 1m prize after answering a series of quick-fire general knowledge questions. It wasn't a close call either – at the final score, Watson's total was 31,547 ahead of its rivals' combined. But in the five years since, IBM's supercomputer has been working towards another goal, one far more lucrative than the Jeopardy!

Summary: I describe how the TrueSkill algorithm works using concepts you're already familiar with. TrueSkill is used on Xbox Live to rank and match players and it serves as a great way to understand how statistical machine learning is actually applied today. I've also created an open source project where I implemented TrueSkill three different times in increasing complexity and capability. In addition, I've created a detailed supplemental math paper that works out equations that I gloss over here. Feel free to jump to sections that look interesting and ignore ones that seem boring. Don't worry if this post seems a bit long, there are lots of pictures. It seemed easy enough: I wanted to create a database to track the skill levels of my coworkers in chess and foosball. I already knew that I wasn't very good at foosball and would bring down better players. I was curious if an algorithm could do a better job at creating well-balanced matches. I also wanted to see if I was improving at chess. I knew I needed to have an easy way to collect results from everyone and then use an algorithm that would keep getting better with more data. I was looking for a way to compress all that data and distill it down to some simple knowledge of how skilled people are. Based on some previous things that I had heard about, this seemed like a good fit for "machine learning." Machine learning is a hot area in Computer Science-- but it's intimidating. Like most subjects, there's a lot to learn to be an expert in the field. I didn't need to go very deep; I just needed to understand enough to solve my problem. I found a link to the paper describing the TrueSkill algorithm and I read it several times, but it didn't make sense. It was only 8 pages long, but it seemed beyond my capability to understand.